System and method for generating distortion free images

Abstract

A camera for generating distortion free images and a method thereof is disclosed. The camera includes a plurality of lenses, wherein each of the plurality of lenses has a dedicated sensor. The camera further includes a processor communicatively coupled to the plurality of lenses. The camera further includes a memory communicatively coupled to the processor and having instructions stored thereon, causing the processor, on execution to capture a plurality of images through the plurality of lenses and to generate a single distortion free image from the plurality of images based on a deep learning technique trained using a mapping of each of a plurality of sets of low resolution images generated in one or more environments to an associated distortion free image, wherein one or more low resolution images in each of the plurality of sets are distorted.

What is claimed is: 
     
       1. A method of generating distortion free images, the method comprising:
 capturing a plurality of images through a plurality of lenses in a camera, wherein each of the plurality of lenses has a dedicated sensor; and 
 generating, by the camera, a single distortion free image from the plurality of images based on a deep learning technique trained using a mapping of each of a plurality of sets of low resolution images generated in one or more environments to an associated distortion free image, wherein one or more low resolution images in each of the plurality of sets are distorted by the plurality of lenses, based on mapping and an environment in which the plurality of images is captured, wherein a distorted pixel is assigned zero weight and wherein generating the single distortion free image comprises the processor to assign weights, using the deep learning technique, to each pixel within each of the plurality of images captured, and wherein the weights to each pixel is assigned by the processor based on:
 determining an environment associated with each image from the plurality of images; and 
 assigning a weight to each pixel of each image from the plurality of images, based on a mapping created between the environment and a corresponding weight. 
 
 
     
     
       2. The method of  claim 1 , wherein generating comprises identifying at least one distorted image from the plurality of images using the deep learning technique, wherein each of the at least one distorted image comprises a plurality of distorted pixels. 
     
     
       3. The method of  claim 2 , wherein generating further comprises:
 discarding the at least one distorted image; and 
 combining a remaining set of images in the plurality of images to generate the single distortion free image, wherein the remaining set of images is obtained after discarding the at least one distorted image. 
 
     
     
       4. The method of  claim 2 , wherein each pixel in each image within a set from the plurality of sets of low resolution images is mapped to a corresponding pixel of the associated distortion free image in the mapping. 
     
     
       5. The method of  claim 4 , further comprising determining a total weight for each of the plurality of images by computing an average of weights assigned to associated pixels. 
     
     
       6. The method of  claim 5 , further comprising combining the plurality of images based on associated total weights to generate the single distortion free image. 
     
     
       7. The method of  claim 1 , further comprising:
 discarding each distorted pixel from each of the plurality of images to generate a plurality of modified images; and 
 combining the plurality of modified images obtained after discarding each distorted pixel from each of the plurality of images to generate the single distortion free image. 
 
     
     
       8. The method of  claim 1 , wherein each of the plurality of images are low resolution images and the single distortion free image is a high resolution image. 
     
     
       9. The method of  claim 1 , wherein the one or more environments
 comprise at least one of rain, fog, dust, lighting conditions, landscape, or an obstacle. 
 
     
     
       10. The method of  claim 1 , wherein the plurality of lenses is arranged as a lens array in the camera. 
     
     
       11. The method of  claim 1 , wherein the one or more environments are simulated to generate the plurality of sets of low resolution images used to create the mapping. 
     
     
       12. A camera comprising:
 a plurality of lenses, wherein each of the plurality of lenses has a dedicated sensor; 
 a processor communicatively coupled to the plurality of lenses; and 
 a memory communicatively coupled to the processor and having instructions stored thereon, causing the processor, on execution to:
 capture a plurality of images through the plurality of lenses; and 
 generate a single distortion free image from the plurality of images based on a deep learning technique trained using a mapping of each of a plurality of sets of low resolution images generated in one or more environments to an associated distortion free image, wherein one or more low resolution images in each of the plurality of sets are distorted and wherein generating the single distortion free image comprises the processor to assign weights, using the deep learning technique, to each pixel within each of the plurality of images captured by the plurality of lenses, based on mapping and an environment in which the plurality of images is captured, wherein a distorted pixel is assigned zero weight, and wherein the weights to each pixel is assigned by the processor based on:
 determining an environment associated with each image from the plurality of images; and 
 assigning a weight to each pixel of each image from the plurality of images, based on a mapping created between the environment and a corresponding weight. 
 
 
 
     
     
       13. The camera of  claim 12 , wherein processor instructions further cause the processor to identify at least one distorted image from the plurality of images using the deep learning technique, wherein each of the at least one distorted image comprises a plurality of distorted pixels. 
     
     
       14. The camera of  claim 12 , wherein processor instructions further cause the processor to:
 discard the at least one distorted image; and 
 combine a remaining set of images in the plurality of images to generate the single distortion free image, wherein the remaining set of images is obtained after discarding the at least one distorted image. 
 
     
     
       15. The camera of  claim 12 , wherein each pixel in each image within a set from the plurality of sets of low resolution images is mapped to a corresponding pixel of the associated distortion free image in the mapping. 
     
     
       16. The camera of  claim 12 , wherein processor instructions further cause the processor to determine a total weight for each of the plurality of images by computing an average of weights assigned to associated pixels. 
     
     
       17. The camera of  claim 16 , wherein processor instructions further cause the processor to combine the plurality of images based on associated total weights to generate the single distortion free image. 
     
     
       18. The camera of  claim 12 , wherein processor instructions further cause the processor to:
 discard each distorted pixel from each of the plurality of images to generate a plurality of modified images; and 
 combine the plurality of modified images obtained after discarding each distorted pixel from each of the plurality of images to generate the single distortion free image. 
 
     
     
       19. A non-transitory computer-readable storage medium that has a set of computer-executable instructions stored thereon, the set of computer-executable instructions causing a camera that includes one or more processors to perform steps including:
 capturing a plurality of images through a plurality of lenses in the camera, wherein each of the plurality of lenses has a dedicated sensor; and 
 generating a single distortion free image from the plurality of images based on a deep learning technique trained using a mapping of each of a plurality of sets of low resolution images generated in one or more environments to an associated distortion free image, wherein one or more low resolution images in each of the plurality of sets are distorted and wherein generating the single distortion free image comprises the processor to assign weights, using the deep learning technique, to each pixel within each of the plurality of images captured by the plurality of lenses, based on mapping and an environment in which the plurality of images is captured, wherein a distorted pixel is assigned zero weight and wherein the weights to each pixel is assigned by the processor based on:
 determining an environment associated with each image from the plurality of images; and 
 assigning a weight to each pixel of each image from the plurality of images, based on a mapping created between the environment and a corresponding weight.